Lobster Roll

All Programming (General) 14276 Systems / Low-Level / OS 11778 Programming Languages / CS Theory 7563 Data / Databases / Infrastructure 6478 Web Development 4631 Security / Privacy 3776 Culture / Philosophy / History / Reading 3691 AI / Machine Learning 3187 Productivity / Career / Business 2564 Maker / DIY / Hardware 2341 Gaming / Retro Computing 1526 Design / UX / Visualization 1503 Science / Math / Physics 1406 Apple / macOS / iOS 1217 Internet / Digital Culture 555 Health / Fitness / Lifestyle 33 Crypto / Blockchain 31

AI / Machine Learning

Where I'm at with AI (paulosman.me)
Reliable Signals of Honest Intent (zanlib.dev)
Anyone scanning published dependencies/packages with AI?
Marking an Exam Done by AI (youtube.com)
My week with opencode (deadsimpletech.com)
AI Is still making code worse: A new CMU study confirms (2025) (blog.robbowley.net)
nanolang: A tiny experimental language designed to be targeted by coding LLMs (github.com)
I was a top 0.01% Cursor user. Here's why I switched to Claude Code 2.0 (blog.silennai.com)
LLMs and your career (notes.eatonphil.com)
XPER on the Commodore 64 (stonetools.ghost.io)
My thoughts on Gas Town after 10,000 hours of Claude Code (simonhartcher.com)
Claude voice mode is still a joke in 2026 (simonhartcher.com)
If Not Lessons, Then What? (third-bit.com)
On the Coming Industrialisation of Exploit Generation with LLMs (sean.heelan.io)
CROW: CROBOTS robotic combat for training World Model AIs (github.com)
Agent Psychosis: Are We Going Insane? (lucumr.pocoo.org)
VaultGemma: A differentially private LLM (2025) (arxiv.org)
“You Had One Job”: Why Twenty Years of DevOps Has Failed to Do it (honeycomb.io)
If writing the code is the easy part, why would I want someone else to write it? (tldraw.dev)
AI Plays Rollercoaster Tycoon (labs.ramp.com)
GitHub Copilot now supports OpenCode (github.blog)
Cursor's latest "browser experiment" implied success without evidence (embedding-shapes.github.io)
Cutting LLM token Usage by ~80% using REPL driven document analysis (yogthos.net)
Weird Generalization and Inductive Backdoors: New Ways to Corrupt LLMs (arxiv.org)
LLM problems observed in humans (embd.cc)
Histomat of F/OSS: We should reclaim LLMs, not reject them (writings.hongminhee.org)
Open Responses Specification (openresponses.org)
My Projects in 2025 (simonhartcher.com)
Sampling at negative temperature (cavendishlabs.org)
When Models Manipulate Manifolds: The Geometry of a Counting Task (transformer-circuits.pub)